The present invention is described in the context of a fluoroscopic imaging system which optionally may use a cardiac gating signal to select certain images from its image stream based on cardiac-cycle data.
Anatomical mapping systems provide the 3D position of a navigational catheter within the cardiac chamber of interest and, in some instances, can also be used to construct 3D maps of the cardiac chamber. These systems are, however, quite expensive to both acquire and operate. Therefore, these systems are available only in a few laboratories for use during interventional procedures, and some of these systems may require specifically-designed catheters such as catheters with built-in sensors.
Conventional fluoroscopy systems are available in all cardiac interventional labs for imaging and real-time navigation of catheters and other instruments and for the placement of leads and stents during interventional procedures. Other than the initial acquisition cost, such systems require little ongoing operation costs. Further, conventional fluoroscopic systems are able to visualize any type of catheter.
FIGS. 1A and 1B illustrate two examples of images obtained from a conventional fluoroscopic system during an atrial fibrillation ablation procedure. Shown in FIGS. 1A and 1B are a mapping and ablation catheter 2, an esophageal probe 3 which is placed inside the esophagus (posterior to the heart), a multi-electrode basket catheter 4, and a coronary sinus catheter 5. These catheters include radio-absorbing material and provide good image contrast compared to the biological tissue such as the lungs 6 and the cardiac silhouette 7. The attenuation of the X-ray beam in the lung is inferior to that of the heart because the lungs are filled with air, the density of which is less than regular anatomical tissue. There is no depth (z-axis) information of the different catheters discernible in these images despite the location of various structures and catheters in different orientations.
As illustrated in FIGS. 1A and 1B, fluoroscopy images produced by conventional systems have the limitation that they do not provide 3D image data. Biplane fluoroscopy (two 2D views from different directions) can be used to identify the relative position of an object such as a catheter. However, its usefulness is limited due to cost and to excessive radiation, and only 1 to 2% of interventional labs have the capability to perform biplane fluoroscopy.
FIG. 2 illustrates a conventional fluoroscopic system 10 used to acquire 2D fluoroscopic image data. The imaging process for conventional fluoroscopy involves an X-ray source 11 which sends an X-ray beam through a patient (not shown) on a table 12. X-ray generation is initiated by actuating a foot pedal 9 on a control panel 15 which is connected (connection not shown) to fluoroscopic system 10. An X-ray detector 13, which may be a flat-panel detector or an image intensifier/video camera assembly, receives the X-rays transmitted through the patient and converts the X-ray energy into an image. X-ray source 11 and X-ray detector 13 are mounted on opposite ends of a C-arm 8. Detector 13 performs the conversion using an X-ray detection layer that either produces light or releases electrons when stimulated by X-rays, and a light-to-electron conversion layer, e.g., photodiodes or electron collection layer, as appropriate, in which an electrical charge signal proportional to X-ray signal intensity in each picture element (pixel) is collected. Analog-to-digital conversion then produces a digital image. Whatever the X-ray detector, the resulting digital image is then processed, possibly stored, and displayed on a screen 14. A control panel is shown at 15. Images may then be displayed on a computer display 14.
FIG. 3 illustrates a coordinate system for fluoroscopic system 10. The z-axis is defined from X-ray source 11 to the center of X-ray detector 13. X-ray (used interchangeably with fluoroscopy) table 12 defines an x-axis and a y-axis. The three axes are shown by the solid lines. The intersection of the axes is the center, or origin O, at (0,0,0) of the 3D space defined by axes x, y and z. Since C-arm 8 is able to move, the z-axis is defined herein when C-arm 8 is oriented in a vertical position as shown in FIG. 2, or a PA position (posterior-anterior position).
X-ray source 11 includes a cathode and an anode. Electrons interact with the anode material producing X-ray photons forming a cone-shaped beam. The beam is controlled by collimator blades to limit the patient's radiation exposure. The X-ray photons propagate in straight lines, forming an image on X-ray detector 13 at a precise location representing the matter encountered all along the ray that goes from the point of X-ray emission from source 11 to this location (pixel in the image) in detector 13. The intensity of the pixel is defined by the type and amount of material (tissue, contrast agent, interventional tool) encountered along this path. The attenuation of the X-ray beam varies as a function of the atomic number and the density of the traversed tissue. The current commercial standard for X-ray image resolution is about 0.2 mm×0.2 mm.
Fluoroscopic images, because they are projections, are representations of the imaged volume of 3D anatomy. This volume is transformed into a 2D projected image on X-ray detector 13 according to precise geometric rules based on the position of X-ray source 11, patient anatomy and detector position in the z direction (parallel to the central ray passing through center O). X-ray projection imaging thus embodies an inherent distortion due to the fact that X-ray source 11 is a finite distance from the anatomy being imaged. As a result, objects closer to X-ray source 11 are magnified in the detected image more than objects more distant from X-ray source 11, and there is no way to resolve these ambiguities without knowing the positions (or a priori sizes) of the objects of interest along the z-axis.
FIG. 3B is an illustration of the geometric magnification which results from the geometric arrangement and conic form of the output of the X-ray source 11 of the X-ray machine 10. FIG. 3B shows a simple 2D representation of the geometry of X-ray machine 10, including source 11, table 12 and detector 13. An object having width wo forms an image having an image width wi on detector 13. (For simplicity, such object and image are also referred to with reference identifiers wo and wi, respectively.) Object wo is located at a distance SOD (source-to-object distance) from source 11, and detector 13 is located at a distance SID (source-to-image distance) from source 11. By simple geometry, the ratio of wi to wo is equal to the ratio of SID to SOD. Thus, the geometric magnification M of such an arrangement is M=SID/SOD.
In U.S. patent application Ser. No. 12/885,710, entitled “3D Model Creation of Anatomic Structures Using Single-Plane Fluoroscopy,” an algorithm for estimating 3D coordinates of a point of interest such as a catheter tip using single-plane fluoroscopy is disclosed. The algorithm computes 3D position estimates by: (1) determining the 3D coordinates of an initial catheter-tip position; (2) advancing the catheter a small, measured amount and obtaining a fluoroscopic image; (3) measuring changes in the position of the catheter tip between the position in the image and the initial position of the catheter tip; (4) computing the actual (physical) changes in catheter-tip position in the x and y directions; (5) computing the 3D position of the catheter tip based on the geometry of the fluoroscopic system and changes in catheter-tip position in x and y; and (6) repeating these steps to generate a series of 3D positions of the catheter tip. This algorithm depends knowledge of the initial 3D coordinates of the catheter tip and advancing the catheter by small, measurable amounts to be able to assume that the catheter tip moves in straight lines. This assumption permits the use of a linear distance formula to compute the 3D coordinates of successive point-of-interest locations. Initial phantom tests have suggested the existing algorithm is reasonable (error <8 mm) when the assumptions are met. However, in some scenarios, the restrictions on catheter advancement are too severe to be able to reduce 3D position errors.
A paper by Pascal Fallavollita entitled “Is Single-View Fluoroscopy Sufficient in Guiding Cardiac Ablation Procedures?” published in the International Journal of Biomedical Imaging, Vol. 2010, Article ID 631264, describes a system which uses X-ray system geometry and image filtering and pattern recognition techniques to estimate the depth (z-axis coordinate) of a catheter tip. The present invention is a significant improvement over the approach of Fallavollita, achieving increased accuracy and doing so in an automated fashion to avoid placing additional requirements on the clinician. The invention identifies the cues present within the fluoroscopic system and uses complex computational algorithms to identify the 3D location of the catheter. Pixel values in fluoroscopic images are affected by out-of-plane angle, depth, and distance from the central ray, and these and other characteristics are taken into account to identify more precise estimates of 3D locations of the catheter tip. With 3D locations of catheter tips determined, various 3D maps such as activation and voltage are created.
It is desired that for a method to be useful, the z-coordinate accuracy of a method of determining 3D coordinates in a fluoroscopic environment should achieve depth (z-coordinate) accuracy of ±4 mm. At least such accuracy is desirable since, for example, a typical lesion formed by a cardiac ablation catheter may be about 4-6 mm in diameter; thus positioning accuracy on that order is desirable for the inventive method to be useful during such interventional procedures.